Four minutes, thirty times a day, is two hours a day. Across a working year that is around four hundred and fifty hours.
Most of a full-time person. Nobody in your business will ever mention it, because four minutes does not feel like anything.
This is the shape of the problem in almost every firm I walk into. Not one big obvious failure. Forty small ones, invisible because they are routine.
You already know the symptoms. Friday afternoons vanish. Quotes go out late. There is always a pile waiting on one particular person. Usually you. What almost no owner can tell me is where the time actually went.
So, the question this article is built around:
Do you actually know what your small tasks add up to?
Not roughly. Not "a fair bit". A number. Until you have one, you cannot tell the difference between a task that is mildly irritating and a task that is costing you a day a week. From the inside those two feel identical.
What follows is the walkthrough I use on a first visit. It takes an afternoon. It costs nothing. You can run it yourself. I would rather you did. An owner who has done this properly can have a sensible conversation about what technology might help. An owner who has not is just buying software and hoping.
Why you cannot see it from the inside
There is a reason you have not spotted it already. It is not that you are not looking hard enough.
When you have done something the same way for years, it stops registering as a task. It becomes part of the furniture. Nobody sits down on a Tuesday and thinks "I spent forty minutes copying numbers from one screen into another." They think "I did the invoicing." The forty minutes vanishes into the label.
Worse, most of these processes were sensible once. Somebody set them up for a good reason. Then the reason went away and the process stayed. A firm I looked at was still producing a detailed weekly report that took the office manager the best part of a morning. It had been introduced years earlier for a single large client who wanted it. That client had left in 2022. Nobody had told her to stop, so she carried on, every week, for a customer who no longer existed.
That is what parked and forgotten looks like. It is not incompetence. It is what happens when work becomes routine and nobody has a reason to look at it again. The people doing it are usually the least likely to question it, because to them it is simply Wednesday.
So you are not looking for the thing that annoys you most. You are looking for the thing you have stopped noticing.
The arithmetic nobody does
Back to that four minute task, because the comparison is the point.
A job that takes ninety minutes and happens once a month costs you eighteen hours a year. You would notice that. You would probably moan about it in the pub.
The four minute job that happens thirty times a day costs you twenty-five times as much. Nobody moans about it at all.
That is the trap. We judge work by how it feels in the moment, not by what it totals. The four minute job feels free. It is the most expensive thing in the building.
This is also exactly the shape of work that is now automatable. Long, complicated, judgement-heavy work is hard to hand over. Short, repetitive, rules-based work is not. So the tasks you are least likely to notice are the ones most likely to be worth doing something about, which is a slightly uncomfortable thought and the reason this exercise is worth an afternoon.
Step one: track a week badly, then note which week it was
Start here. Do not overthink it.
For one week, every person in the business writes down what they did and roughly how long it took. Not a formal time sheet. A notebook, a note on the phone, whatever they will actually use. Fifteen minute chunks are fine. Rough is fine. If someone writes "chased suppliers, ages" that is still useful.
Two rules make this work.
First, nobody is being assessed. This matters more than anything else in this article. The moment a member of staff thinks this is a productivity audit, the data becomes worthless. They will tidy it up. They will leave out the bits they think make them look slow. You will end up with a beautiful record of a week that did not happen. Say plainly, at the start: we are looking at the work, not the worker. If something took two hours because the system is daft, write two hours. That is the useful bit.
Second, the part I see people get wrong most often: write down which week it was.
Because your week in February is not your week in July. Forget that and you will draw the wrong conclusions from good data.
Almost every business runs on a rhythm. A landscaper's problem in April is the diary and the weather. In January it is cash and quoting for spring. A retailer's November has nothing in common with their February. An accountancy practice in the run-up to the end of January is a different organisation to the same practice in June. A care operator's summer is annual leave cover and agency staff. A wholesaler has the pre-Christmas surge and then a long flat stretch. Many trades and service firms have a monthly cycle too, with everything piling into the last week when invoicing and payroll land on top of the normal work.
So when you look at your notes, ask of each item: does this happen all year, or is this a peak thing?
The all-year tasks are your baseline. Those are where automation pays back most reliably, because the saving repeats fifty-two weeks a year.
The seasonal ones are worth flagging separately. They matter for a different reason. A task that is manageable in a quiet month can be the exact thing that breaks you in a busy one. Plenty of firms cope fine for nine months and then hit a wall. The wall is almost always an admin process that was fine at half the volume.
Two practical suggestions. Pick a week that is genuinely typical for the time of year, not the week of the big job and not the week half the team is off. And if you can, run this twice, once now and once at the opposite end of your cycle. The comparison is often more revealing than either week on its own, because it shows you what actually scales and what quietly falls over.
If you can only do it once, do it once. A single messy real week beats a tidy imaginary one. It beats not doing it at all by a distance.
Step two: sort what you find into four piles
At the end of the week you will have a pile of scrappy notes. Now sort every task into one of four categories. Do this on paper, on a wall, wherever. It should take an hour or two.
Pile one: the work itself. Fixing the boiler. Seeing the patient. Making the thing the customer pays for. This is the point of the business.
Pile two: work that helps the work. Planning, ordering, scheduling, training, quoting. This is not the product, but it is not waste either. A business without it falls over.
Pile three: moving information around. Copying a number from one place to another. Retyping something a customer already told you. Sending an email to say an email arrived. Reading out a job sheet down the phone so someone else can write it down again. This is the pile that matters.
Pile four: fixing avoidable mistakes. Redoing something because the information was wrong, late, or lost. Chasing something that should not have needed chasing.
Most owners are genuinely shocked by piles three and four. Not because they are enormous, but because until you write them down they never appear anywhere. They are not on a job sheet. They do not have a cost code. They just happen, in the gaps, all day.
Now do the arithmetic on pile three. Multiply each item by how often it happens, across a year, allowing for the seasonal ones being heavier at peak. That total is the number you did not have at the start of this article.
Step three: ask the two questions that actually find the bottleneck
Go through pile three and pile four and ask two things about each item.
"How often does this happen? At what time of year?"
Frequency beats duration almost every time. It is also the thing people consistently get wrong. We remember the big annoying job and forget the small constant one. Add the timing question on top: something that happens daily all year is a different proposition to something that happens daily for eleven weeks and then stops.
"What is waiting while this happens?"
This is the question that finds the real bottleneck. Hardly anyone asks it.
Some tasks are slow but nothing waits for them. You do the VAT return in the evening and no customer is sitting on their hands while you do it. Annoying, but not a bottleneck.
Other tasks are quick but everything queues behind them. Say quotes only go out when you personally sit down to write them. Say you only get to that on a Thursday. Every enquiry from Friday to Wednesday is then sat waiting on you. The task might be fifteen minutes. The delay is six days. Somewhere in that six days a customer rings somebody else.
That gap, between how long something takes and how long it holds things up, is where the money is. A bottleneck is not the slowest task. It is the one everything else is queuing behind.
Step four: score them honestly
Take your top ten and give each one two scores out of five.
Pain: how much time it costs across the year, plus how much it holds up.
Difficulty: how hard it would be to change. Be honest here rather than optimistic. Does it need new software, or just a decision? Does it touch a customer, or is it purely internal? Would it need three people to change how they work, or one?
Then start with high pain and low difficulty. Always. Not the most exciting one, not the one a salesperson got you interested in. The dull, easy, expensive one.
The reason is not really about efficiency. It is that the first change has to work. If your first attempt is the ambitious one and it fails, you will not get a second go, because nobody in the building will believe the next idea either. Win the easy one first and buy yourself the credibility to try the hard one.
What usually turns up
I have run this in field service firms, care operators, professional practices, wholesalers and kitchens. Different industries, same list. Nearly every time, these four are in the top five.
Information typed twice. A customer tells you something. It goes on a form. Someone types the form into the system. Someone else types it into the invoice. Same facts, three journeys, three chances to get it wrong. The engineer who fills in a paper job sheet on site and then retypes it into the system that evening is doing the job twice. He is doing the second half in his own time.
Somebody being a human search engine. One person knows where everything is, what the price was last time, which customer likes it done a particular way, which resident does not like being woken early. This one is chronic in field service and trades, where the knowledge lives in one engineer's head and travels around in his van. When that person is off, the business slows to a crawl. Everyone treats this as loyalty. It is actually a single point of failure. It gets exposed at exactly the wrong moment, usually in the busiest month.
Chasing. Chasing suppliers, chasing engineers, chasing clients for the documents you need to do the work they asked you to do. An accountancy practice I know reckoned chasing paperwork was the single largest use of junior time in January. Chasing is close to pure waste. It is usually a symptom of information not moving on its own.
Scheduling by argument. The daily rota, the job list, who goes where, who covers the shift. Anyone in hospitality or catering will recognise this one before they finish reading the sentence. Often solved in a twenty minute conversation every morning that could have been a two minute glance at something everyone can already see.
If you finish this exercise and none of those appear, either you have already sorted them, which is genuinely rare, or the week you tracked was not a typical one.
Only now, think about technology
Here is the part where I am supposed to tell you AI will fix all of it. It will not. Anyone promising that has not looked at your business.
What is fair to say is that a good chunk of pile three, moving information around, is now automatable in a way it simply was not five years ago. Reading a supplier invoice and pulling the numbers off it. Turning an engineer's spoken notes into a written job report before he has left the car park. Drafting a first version of a quote from a phone call. Watching for the thing that has not happened yet and flagging it before it becomes a problem. These are real. They work. They are not expensive any more.
But notice the order. You find the bottleneck first. You work out what it costs you across a year, peaks included. You understand what is queuing behind it. Then you ask what could take that job on. The answer is sometimes software, sometimes a different process, quite often just a decision that one person no longer has to be the gate everything passes through.
The failures I see are almost always the other way round. Somebody buys a system, then goes looking for a problem it solves. The system usually works exactly as advertised. It just automates something that was never the bottleneck, so nothing gets faster and everyone concludes that technology does not work here.
One more thing worth saying plainly, because it is the question underneath the question. In practically every one of these I have run, the outcome has not been fewer people. It has been the same people doing more of pile one and less of pile three. It has been coping with the busy season without hiring in temporary help to manage the paperwork. That is the honest version of what this does. It is a better story than the scary one.
If you want a hand
Run it yourself. Genuinely. Most owners who do this find at least one thing worth fixing without any outside help at all. You will understand your own business better for having done it.
If you get through it and you are looking at a list you are not sure how to act on, that is what we do.
The first conversation is free. Always. No pitch. No obligation. No expectation that it leads anywhere. Bring your four piles and we will go through them with you. If the honest answer is that you do not need us, you will hear that. It will not have cost you anything to find out.
If it does make sense to go further, the next step is a formal efficiency analysis: properly, across the whole business, across the year rather than a single week. We tell you plainly what is worth automating, what is not, plus what it would actually cost before you commit to anything. Sometimes the answer is that your bottleneck is a process problem and no software will touch it. We will tell you that too.
And if the thing standing between you and better use of the tools you already pay for is simply that nobody has been shown how, that is a different job. We train teams to use AI properly as well.
Either way, start with the week. Write down which week it was. Then do the arithmetic, because you cannot fix what you cannot see. You will not see it until you turn it into a number.
Next Monday: Making Tax Digital started on 6 April. The first quarterly deadline is 7 August. What around 864,000 sole traders and landlords must actually do, by when, plus what happens if they miss it.